Genetic Variation in Artemisia Campestris

Genetic variation in Artemisia

campestris: do distinct genetic varieties exist within the species? United States Department of Agriculture Lab Report

NFGEL Project #319

Forest Service Project submitted by Andrew Bower, Area Geneticist Mark Skinner, Regional Botanist US Forest Service US Forest Service 1835 Black Lake Blvd. SW Region 6 Regional Office Olympia, WA 98512 1220 SW 3rd Avenue p: 360-956-2405 Portland, OR 97204 Washington Office [email protected] p: 503-808-2150 Forest Management [email protected] Theodore B. Thomas, Ecologist Pacific Region, Washington Fish and Wildlife Office Branch of Listing, Critical Habitat 510 Desmond Drive SE, suite 102 Lacey, WA 98501 p: 360-753-4327 National Forest [email protected] Genetics Laboratory (NFGEL) This work was conducted under the direction of

Valerie Hipkins, PhD……...…………………………..NFGEL Director

Report prepared by

Barbara Wilson, PhD. Carex Working Group, Corvallis, Oregon 2480 Carson Rd Placerville, CA Laboratory technical work conducted by 95667 Andrew Jackson, Keenan Raleigh, Valerie Hipkins, 530-622-1609 Tissue Preparation Ed Sprigg, Amanda Cutler, Jennifer DeWoody, Jacob (phone) Snelling Ploidy Analysis Keenan Raleigh, Amanda Cutler, Valerie Hipkins 530-622-2633 (fax) Randy Meyer, Amanda Cutler, Jian Alsarraj, Valerie Isozyme Analysis [email protected] Hipkins

Valerie Hipkins May 7, 2016 NFGEL Artemisia Report

INTRODUCTION The pattern of variation in certain herbaceous Artemisia campestris along the Columbia River is complex and does not lend itself to simple taxonomic classification. In the Flora of the Pacific Northwest, Artemisia borealis is treated as a subspecies of A. campestris, and A. campestris is described as “A highly polymorphic species, divisible into a number of rather diverse but apparently wholly confluent subordinate taxa” (Hitchcock et al. 1955, p. 60). The taxonomic history of Artemisia campestris and A. borealis is more like a braided stream than the expected tree (Arnett 2012) and taxonomists disagree about how to treat the taxa (Table 1).

To add to the confusion, both A. borealis and A. campestris are variable. Two forms of A. borealis and/or A. campestris occur along the Columbia River in Washington and, historically, in Oregon. One form, “wormskioldii,” is typically biennial and is short, with hairier leaves and stems, slightly broader leaves, and more compact inflorescences with bigger flowers. It blooms in April and May with occasional stragglers blooming into later summer. The other form, “scouleriana,” is usually perennial and is much taller, though heights overlap, with less hairy leaves and stems, narrower leaves, and long, open inflorescences with smaller flowers. It blooms in August and September. The wormskioldii plants resemble arctic/alpine A. borealis (Figures 1 and 2). The scouleriana plants resemble widespread low elevation A. campestris. Intermediates occur, and seeds collected from wormskioldii plants occasionally produce an offspring with the scouleriana phenotype (K. Amsberry, per. comm.). The wormskioldii form lives on sand dunes and similar habitats along the Columbia. This habitat has been mostly destroyed as a consequence of dam-building, and two wormskioldii populations remain, at Beverly and at Miller Island. The scouleriana form is widespread in the Pacific Northwest. It is fairly common at several sites along the Columbia River, including Beverly and Miller Island.

The wormskioldii form is a listed rare plant; critically imperiled globally, endangered in Washington, and a candidate species with the USFWS (Washington Natural Heritage Program 2014). It is heavily managed and new populations have been established. However, the mixed populations at Beverly and Miller Island, and the existence of morphologically intermediate plants raise the question of whether wormskioldii is a real taxon or a set of extreme phenotypes in variable scouleriana populations. If wormskioldii is a not a real taxon, then it could be delisted and resources used to preserve it could be re-directed to manage other rare plants.

This isozyme study was initiated to assess the genetic variation within and among wormskioldii and scouleriana occurrences to determine if these forms are genetically distinct. The evolutionary relationships of both of these forms were also compared to A. borealis samples. For purposes of this report, the three putative taxa tested are identified as A. borealis, A. campestris var. scouleriana, and A. campestris var. wormskioldii.

NFGEL Artemisia Report

METHODS

SAMPLE COLLECTION

A total of 310 samples of Artemisia spp. were submitted for this project, with 31 plants collected from each of 10 populations, all in Washington (Table 2; Figure 3). Two populations from the Columbia River were identified as A. campestris var. wormskioldii, two populations from the Olympic Mountains were identified as A. borealis, and the remaining populations were identified as A. campestris var. scouleriana. Most of the scouleriana populations were collected on the Columbia River from Beverly to Miller Islands, inclusive, and two came from Whidbey Island in Puget Sound. Samples were collected into individual plastic bags and kept cool until shipment to NFGEL using overnight delivery.

PLOIDY ANALYSIS

Flow cytometry was used to compare relative DNA content among a subset of samples from each population. An approximately 1.0 cm2 piece of leaf tissue was diced in 0.4 mL of Partec® CyStain extraction buffer using a double-edge blade for approximately 30 seconds. Immediately following, 1.6 ml of the CyStain staining buffer was added to the sample, which was then placed through a 10 µm filter and immediately analyzed on a Partec PA Ploidy Analysis system using the parameters: gain = 380, LL = 30, speed = 3, linear scale.

ISOZYME ANALYSIS

All samples were prepared for isozyme analysis by grinding an approximately 2 cm2 piece of leaf tissue with a mortar and pestle under liquid nitrogen until it was a fine powder, then adding around 400 µL (12-15 drops with a plastic pipette) of Tris based extraction buffer (Cheliak and Pitel 1984) and allowing the slurry to freeze to the mortar. When the slurry thawed, 100 µL was transferred to each of three flat-bottom 96-well plates, which were then frozen at -80°C until analysis.

Starch gel (11% v/v) electrophoresis was used to examine isozyme variation. Three buffer combinations from Conkle et al. (1982) were used to examine 16 enzyme stains using recipes adapted from Wendel & Weeden (1989), revealing a total of 9 loci used for analyses. Five loci were examined in a lithium borate electrode buffer-tris citrate gel buffer combination (system LB): fluorescent esterase (FEST; EC 3.1.1.-), leucine aminopeptidase (LAP; EC 3.1.11.1), phosphoglucomutase (PGM; EC 5.4.22.-), malic enzyme (ME7; EC 1.1.1.40), and aconitase (ACO; EC 4.2.1.3). Seven loci were examined in a sodium borate electrode buffer-tris citrate gel buffer combination (system SB): aspartate catalase (CAT; EC 1.11.1.6), triose-phosphate isomerase (TPI; EC 5.3.1.1), uridine diphosphoglucose pyrophosphorylase (UGPP; EC 2.7.7.9), phosphoglucose isomerase (PGI; EC 5.3.1.9), aminotransferase (AAT; EC 2.6.1.1), glycerol dehydrogenase (GLYDH; EC 1.1.1.6), and 6-phosphogluconic dehydrogenase (6-phospho-D- gluconate) (6pgd; EC 1.1.1.44). Five loci were examined on a morpholine citrate electrode and gel buffer (system MC6): malate dehydrogenase (MDH; EC 1.1.1.37), diaphorase (DIA; EC 1.6.99.-), isocitrate dehydrogenase (NADP form) (IDH; EC 1.1.1.42), shikimate dehydrogenase

NFGEL Artemisia Report

(SKD; EC 1.1.1.25), and 6-phosphogluconic dehydrogenase (6-phospho-D-gluconate) (6pgd; EC 1.1.1.44).

Three of these enzyme stains were run and found to be monomorphic for all samples (FEST, CAT, AAT-1 and AAT-2); three other stains were found to have insufficient activity among samples (LAP, IDH, and MDH); and one stain (UGPP) had no activity in A. campestris samples but contained two segregating alleles in the A. borealis samples. 6PGD was run on both the SB and MC6 buffer systems, but only data from the SB system was used in the analysis. Data from nine isozyme loci were used in the final analysis.

DATA ANALYSIS

Isozyme data were examined for standard measures of genetic diversity within populations (alleles per locus, effective alleles per locus, observed heterozygosity, and fixation) and for differentiation among populations using GenAlEx v. 6 (Peakall and Smouse 2006). If the two putative subtaxa (wormskioldii and scouleriana) represent distinct subspecies or varieties, we predicted that populations of each subtaxa would be more genetically similar to each other than they were to populations from another subtaxa or species. Put another way, two putative taxa are expected to show greater differentiation than individual populations within each taxa. Genetic differentiation among populations was examined using analysis of molecular variance (AMOVA) and principal coordinates analyses (PCoA), both implemented in GenAlEx v. 6.

In addition, Nei’s (1972) unbiased genetic distance between all pairs of populations was visualized using UPGMA to build a population phenogram using Mega v. 5.2 (Tamura et al. 2011). Admixture analyses and population assignment tests were conducted in Structure v. 2.3.4 (Falush et al. 2003, 2007; Pritchard et al. 2000), estimating likelihoods for 500,000 replicates following a 50,000 burn-in period, with parameters set to assume independent allele frequencies among populations, with other parameters set to the default. The number of genetic clusters tested (K) was defined as K={2:8}, and each K was examined five times. The results of the admixture analyses were examined using two methods. First, the mean and variance of the log- likelihood of each K (over the five replicates) was examined to identify the K with the greatest likelihood and low variance. Second, the delta-K method of Earl & vonHoldt (2012) was used to examine the decrease in the rate of change of log-likelihood over each K as a method to choose the most likely number of genetic clusters, as employed by Structure Harvester v. 0.6.94. Experience has shown that for variable outcrossing populations, Structure Harvester may underestimate K, so the visual approach may provide additional information about the genetic structure of the populations.

RESULTS

DNA CONTENT AND PLOIDY

Five samples per population were analyzed on a flow cytometer to estimate relative DNA content. Differing DNA content is indicative of putative ploidy variation among populations which may indicate taxonomic differences among samples. In addition to the 50 individual

NFGEL Artemisia Report sample ploidy runs, multiple samples were run together in combination. All 40 A. campestris samples analyzed had an average peak position at 234. The ten A. borealis samples had an average peak position at 117. Peak position is a measure of DNA content (channel number). When two samples of the same putative taxon were run together, only one high peak was observed, indicating that DNA content was consistent among the borealis samples and among the scouleriana + wormskioldii samples, respectively. DNA content of the scouleriana and wormskioldii samples was twice as much as in the borealis samples. This is consistent with the scouleriana and wormskioldii samples being tetraploid and the borealis samples being diploid.

Chromosome counts reported in IPCN (Table A1) include both diploid (2n = 18) and tetraploid (2n = 36) chromosome numbers for both Artemisia borealis (x = 9) and A. campestris (x = 8). Sometimes a single taxon can include plants with different numbers of chromosomes, although in this case the possibility that some samples were misclassified cannot be ruled out. Both A. borealis and A. campestris are circumboreal and North American populations are underrepresented in this compilation (Table A1).

Isozyme banding patterns from the A. borealis and A. campestris samples in this study could all be scored as diploid. They lacked the fixed heterozygosity and unbalanced heterozygotes characteristic of tetraploid isozyme band patterns. For purposes of analysis, therefore, all the samples were treated as diploid. This simplification may make the borealis samples appear more similar to the scouleriana + wormskioldii samples than is warranted, but does not affect the main question of this study, the relationship between wormskioldii and scouleriana.

ALLELE FREQUENCIES

Isozyme alleles were similar in all three putative taxa studied (Table A2). For each locus, the same allele was most common in all populations, with two exceptions: PGI-2 and TPI alleles that were most common in all other populations had frequencies of 45-48% in the Beverly population of wormskioldii. For ACO, the borealis samples were monomophic and the scouleriana + wormskioldii samples were not; for GLYDH and 6PGD, the situation was reversed. All three putative taxa had approximately the same percent of polymorphic loci (Table A3), but differed in which loci were polymorphic.

In A. borealis, UGPP was scoreable and polymorphic. In scouleriana + wormskioldii, it produced an unscorable smear under the test conditions. This seems to indicate a fixed difference at this locus between the two putative taxa A. borealis and A. campestris, with wormskioldii sharing the scouleriana form of the enzyme. This difference may be the result of apparent DNA content differences detected between A. borealis and A. campestris samples.

GENETIC STRUCTURE (TAXONOMIC IMPLICATIONS)

Nei’s genetic identities for the various populations sampled were high, varying from 0.890 to 0.992 (Table 3). A value of 1.000 indicates a population pair that shares all alleles in the same frequency and is therefore genetically identical. The lowest identity was between an A. borealis population high in the Olympic Mountains and a scouleriana population on Whidbey Island in Puget Sound. The highest genetic identity between pairs of populations was between the

NFGEL Artemisia Report

Beverly populations of scouleriana and wormskioldii. A phenogram of these results suggests that the most divergent populations are the pair collected on the shore of Whidbey Island in the Puget Sound, and in general the populations cluster into groups by location (Figure 4).

Geographic distances among populations varied from 0, for the paired scouleriana and wormskioldii populations at Miller Island and at Beverly, to 313.4 km between the Hanford Reach of the Columbia River and Whidbey Island in Puget Sound (Table A4). Genetic distances also varied (Table A5) from the smallest distance of 0.008 (between the two populations located at Beverly) to the greatest distance of 0.117 between the Ebey’s Landing population of scouleriana on Whidbey Island and the Blue Mountains A. borealis population.

The relationship between genetic and geographic distances was tested (Table A6). There is a low but probable correlation between the two when all samples are compared (r = 0.277, p = 0.02; Figure 5). However, the correlation is much higher when the borealis is removed from the comparison (r = 0.612, p = 0.04; Figure 6). Removing the borealis samples may be appropriate because they differ in their high elevation habitat (Table 2) as well as their putative taxonomic difference.

Analysis of molecular variance reveals that about 75% of the genetic variation detected in this study occurs within populations. About 5% of the variation occurs among putative taxa (scouleriana, wormskioldii, and borealis), and comparatively, 20% of the variation is found among locations. In other words, putative taxa explain a small percentage of the variation in these plants, while location explains a much greater percentage of the genetic variation.

Principal component analysis (PCA) based on genetic distances among populations (Figure 7) separated the high elevation borealis populations and the Puget Sound scouleriana populations from the Columbia River populations of wormskioldii + scouleriana. The scouleriana and wormskioldii populations along the Columbia River clustered together, with those of the Hanford Reach intermediate between the Beverly and Miller Island populations. The two axes graphed explain 86% of the variation among populations.

Clustering analysis using Structure revealed the most likely number of genetic groups to be either three or four (Figure 8). The individual assignment tests reveal high levels of genetic mixing, and genetic clustering by location is especially strong. In both K=3 (Figure 8A) and K=4 (Figure 8B), the two borealis populations form a genetic group. The remaining groupings strongly follow population location, not taxon: the Beverly populations group together (populations 2 and 4), Whidbey Island populations (populations 7 and 8) cluster together, and the Miller Island populations (populations 1 and 3) fall in a genetically similar group. There is some distinctive genetic clustering of the Hanford populations (5 and 6), especially when K=4, however the Hanford populations also shows some clustering similarity to the Whidbey Island populations (all four are scouleriana).

NFGEL Artemisia Report

DISCUSSION

This study was initiated to test whether wormskioldii populations and scouleriana populations of Artemisia campestris should be treated as distinct taxonomic entities. The isozyme diversity analyzed indicates that there are moderate levels of genetic variation and significant differentiation among the ten populations of Artemisia sampled. The structure of the genetic diversity supports the taxonomic distinction of the A. borealis samples from the A. campestris var. wormskioldii and var. scouleriana samples (Figures 7 and 8). However, these markers do not support the separate taxonomic status of the two putative varieties within A. campestris (Figures 4, 7, and 8). The genetic differentiation was not consistent with taxonomic divergence between the two A. campestris varieties and therefore does not support the idea that wormskioldii and scouleriana are distinct taxa.

Significant differentiation was observed among locations of sampled populations. The two populations growing at a given location were more similar to each other than were populations identified as the same variety. In other words, if the varietal status is valid, we would expect all populations of a variety to be more genetically similar to each other than they are to another variety or taxon. This was not the case and instead genetic clustering occurred by location suggesting that geography explains a greater proportion of the genetic variability than does the a priori varietal designations (Figures 4, 7, and 8). What we found is that plants growing in the same place are more similar to each other than to more distant populations. The scouleriana and wormskioldii samples at Beverly have a genetic identity of 0.992 (Table 3). Those growing on Miller Island have a genetic identity of 0.982. However, the genetic identity of wormskioldii populations growing on Beverly and Miller Island is only 0.944 and the genetic identity of scouleriana populations at these two sites is only 0.951. In general, comparisons of the other scouleriana populations with the Beverly or Miller Island populations of that putative taxon are intermediate (Table 3). More important than absolute numbers is the expected pattern of variation; we would expect members of a single subspecies to be more similar to each other than to members of a different subspecies.

Morphological traits that can also function as genetic markers indicate that the two varieties are different, perhaps with different ancestry, even though intermediate plants exist. The morphological and phenological differences between the wormskioldii and scouleriana plants along the Columbia are often clear, and they ally wormskioldii with borealis, not scouleriana (Figures 1 and 2). These differences are exhibited in plants growing together at the same sites, so they are not just environmentally induced. In other words, these traits are good genetic markers, like the isozymes studied here. Morphology and phenology indicate that wormskioldii and scouleriana are different entities, although the extent in which the wormskioldii phenotype is heritable should be determined through common garden experiments.

However, isozyme variation was not patterned in a way expected if wormskioldii and scouleriana are different taxonomic groups, and was not able to genetically separate wormskioldii from scouleriana. The genetic identity of 0.992 between wormskioldii and scouleriana at Beverly is very high and indicates great gene flow between the two forms. Looking at isozymes alone, we would call these the same taxon. However it should be stressed that the course resolution provided by these genetic markers reveal broad-scale patterns of

NFGEL Artemisia Report differentiation and may not be sufficient to provide more fine-scale resolution. Further studies using alternative DNA-based nuclear or organelle markers may reveal additional information about the varietal relationships.

What could explain this discrepancy between isozyme markers and morphology/phenology markers? Three scenarios seem likely. Perhaps wormskioldii and scouleriana split relatively recently from a common ancestor, and morphology and phenology are changing under selection, but the more neutral enzymes studied here are not. Alternatively, perhaps wormskioldii were once different, but interbreeding with scouleriana is slowly reducing the neutral genetic differences. And finally, it is possible that the environmental conditions on sandy or gravelly sites at the edge of the river select for the wormskioldii phenotype and early flowering, but are neutral about the basic metabolic enzymes studied here. Therefore, isozymes may be indicative of strong gene flow, but morphology and phenology are reflective of strong selection pressures.

The Puget Sound scouleriana are definitely different from wormskioldii, but that may be an indicator of peculiarities of the Puget Sound plants, rather than the relationships among Columbia River scouleriana and wormskioldii. Hitchcock et al. (1955, p. 60) wrote: “A few of the plants from our range, particularly about Puget Sound, are apparently referable to the otherwise more eastern ssp. caudata . . . . The ssp. caudata seems fairly well distinguished from ssp. borealis in c. and c. U.S., but westerward it becomes difficult to distinguish from var. scouleriana of ssp. borealis.” It would be interesting to know about the genotypes of plants upstream of Beverly or away from the main stem of the Columbia, which presumably would not be ssp. caudata.

Morphology strongly suggests that wormskioldii is the same as borealis. However, wormskioldii differs from the borealis samples in DNA content per cell and in isozyme variation (Figures 4, 7, and 8). It is possible, though, that even if wormskioldii is related to borealis, it is not closely related to the borealis populations of the Olympic Peninsula. There is no likely recent path of plant migration between Beverly and the Olympic Mts. However, borealis from the Rocky Mountains could easily have washed down the Columbia River during normal floods or the Bretz floods some 15,000 years ago, and established population on the shores. It is interesting to speculate that the hairiness, short stature, and early maturity of alpine borealis gave such migrants an advantage on sandy or gravelly banks that went dry early in the season.

Although isozyme variation provided no evidence of two distinct genetic varieties within A. campestris, the significant allele frequency differentiation among populations indicated that genetic structure exists among the sampled populations. Morphological, ecological, and genetic divergence between samples is ultimately required for taxonomic divisions (Crawford 2000). The genetic variation detected among populations studied here may be insufficient to restrict or promote taxonomic changes.

NFGEL Artemisia Report

ACKNOWLEDGEMENTS

Personal communication: Kelly Amsberry, Oregon Department of Agriculture, Rare Plant Program. Oregon State University. [email protected]. Interviewed by Barbara Wilson, winter 2015. Draft report was submitted January 21, 2016 to US Fish and Wildlife Service. Draft report review was provided by Andy Bower, US Forest Service, and by Justin Bohling, Christian Smith, and Brice Adams, Abernathy Fish Technology Center, US Fish and Wildlife Service.

REFERENCES

Arnett, J. 2012. Artemisia campestris ssp. borealis var. wormskioldii: Taxonomic review. Natural Heritage Report 2012-02. Washington State Department of Natural Resources, U.S.F.S. Region 1, WA, U.S.A Cheliak,WM and JA Pitel. 1984. Techniques for starch gel electrophoresis of enzymes from forest tree species. Information Rep.PI-X-42. Ottawa:Petawawa Natl. Forestry Inst.,Canadian Forestry Service, Agriculture, Canada Conkle, MT, PD Hodgskiss, LB Nunnally, and SC Hunter. 1982. Starch Gel Electrophoresis of Conifer Seeds: A Laboratory Manual. USDA Forest Service General Technical Report PSW-64. Pacific Southwest Forest and Range Experiment Station, Berkeley, California. Cook, Thea, Stephen C. Meyers and Scott Sundberg, eds. 2015. Checklist of Oregon Asteraceae, version 1.5. Oregon Flora Project. http://www.oregonflora.org/family_treatments/Asteraceae.pdf (Accessed 5 January 2016). Crawford DJ. 2000. Plant Macromolecular Systematics in the Past 50 Years: One View. Taxon 49, 479-501. Earl, Dent A. and vonHoldt, Bridgett M. 2012. STRUCTURE HARVESTER: a website and program for visualizing STRUCTURE output and implementing the Evanno method. Conservation Genetics Resources 4(2): 359-361. doi: 10.1007/s12686-011-9548-7 Falush, Daniel. 2003. Inference of population structure using multilocus genotype data: linked loci and correlated allele frequencies. Genetics164:1567–1587. Falush, Daniel, Mathew Stephens, and Johnathan K. Pritchard. 2007. Inference of Population Structure Using Multilocus Genotype Data: Dominant Markers and Null Alleles. Molecular Ecology Notes 7(4): 574–578. doi:10.1111/j.1471-8286.2007.01758.x. Hitchcock, C. Leo, Arthur Cronquist, Marion Ownbey, and J. W. Thompson. 1955. Vascular Plants of the Pacific Northwest, Part 5: Compositae. University of Washington Press, Seattle, Washington. IPCN; Index to plant chromosome numbers. 1979 and continuously updated. P. Goldblatt & D. E. Johnson, eds. Missouri Botanical Garden, St. Louis. http://www.tropicos.org/Project/IPCN. Accessed December 2015. Nei, M. 1972. Genetic distance between populations. Am. Nat. 106: 283-292. Peakall R, Smouse PE. 2006. GENALEX 6: genetic analysis in Excel. Population genetic software for teaching and research. Molecular Ecology Notes 6, 288-295.

NFGEL Artemisia Report

Pritchard, Jonathan K., Matthew Stephens, and Peter Donnelly. 2000. Inference of Population Structure Using Multilocus Genotype Data. Genetics 155(2): 945-959. Shultz, Leila M. 2006. Artemisia. pp. 503-534 in Flora of North America Editorial Committee, eds., Flora of North America, volume 19, Magnoliophyta: Asteridae, part 6: Asteraceae, part 1. Oxford University Press, New York. Tamura K, Peterson D, Peterson N, Stecher G, Nei M, Kumar S (2011) MEGA5: Molecular Evolutionary Genetics Analysis using maximum likelihood, evolutionary distance, and maximum parsimony methods. Molecular Biology and Evolution 28, 2731-2739. USDA, NRCS. 2016. The PLANTS Database (http://plants.usda.gov, 7 January 2016). National Plant Data Team, Greensboro, NC 27401-4901 USA. (Accessed December 2015) Washington Natural Heritage Program. 2014. Rare Plants. http://www1.dnr.wa.gov/nhp/refdesk/lists/plantrnk.html#key. Accessed January 7, 2016. Wendel, JF and NF Weeden. 1989. Visualization and interpretation of plant isozymes. Pages 5 – 45 In D. E. Soltis and P. S. Soltis (editors), Isozymes in Plant Biology. Dioscorides Press, Portland, Oregon.

NFGEL Artemisia Report

Table 1. Taxonomic synonyms for selected varieties in the Artemisia borealis / campestris complex. (* wormskioldii is not explicitly treated in FNA, but the plants easily key to A. borealis ssp. borealis as defined in that work.)

Genus, species subtaxon subsubtaxon Database Source borealis A. borealis ssp. borealis FNA Shultz 2006 A. campestris ssp. borealis var. borealis Hitchcock et al. 1955 A. borealis ssp. borealis OFP Cook et al. 2015 A. campestris ssp. borealis var. borealis USDA PLANTS USDA, NRCS 2015 scouleriana A. campestris ssp. pacifica FNA Schultz 2006 A. campestris ssp. borealis var. scouleriana Hitchcock et al. 1955 A. campestris var. scouleriana OFP Cook et al. 2015 A. campestris ssp. borealis var. scouleriana USDA PLANTS USDA, NRCS 2015 wormskioldii A. campestris ssp. borealis var. wormskioldii Arnett 2012 A. borealis ssp. borealis* FNA Schultz 2006 A. campestris ssp. borealis var. wormskioldii Hitchcock et al. 1955 A. campestris var. wormskioldii OFP Cook et al. 2015 A. campestris ssp. borealis var. wormskioldii USDA PLANTS USDA, NRCS 2015 A. campestris var. wormskioldii WA DNR

FNA = Flora of North America; OFP = Oregon Flora Project Database; WA DNR = Washington State Department of Natural Resources.

NFGEL Artemisia Report

Table 2. Sampling locations for isozyme study in A. campestris and A. borealis. All locations are in Washington. The Beverly, Hanford, and Miller Island locations are along the Columbia River. The Whidbey Island site is in Puget Sound, and the two mountain sites are in the Olympic Mountains. Elevations are approximate, derived from Google Earth. Latitude and longitude is based on WGS84 projection. The population number is used in other tables and figures to identify the locations sampled.

Elevation Population Location Taxon Latitude Longitude (ft) #

Miller Island A. campestris var. wormskioldii 45.65882 -120.88001 177 1 Beverly A. campestris var. wormskioldii 46.84091 -119.94319 684 2 Miller Island A. campestris var. scouleriana 45.65882 -120.88001 177 3 Beverly A. campestris var. scouleriana 46.84091 -119.94319 684 4 Hanford A. campestris var. scouleriana 46.67363 -119.45378 405 5 Hanford A. campestris var. scouleriana 46.64116 -119.74102 545 6 Whidbey Island A. campestris var. scouleriana 48.15958 -122.68185 7 7 Whidbey Island A. campestris var. scouleriana 48.18765 -122.70259 26 8 Buckhorn Mountain A. borealis 47.81965 -123.13258 6086 9 Blue Mountains A. borealis 47.95470 -123.25950 6422 10

Table 3. Nei’s genetic identities for sampled populations (#1- #10) of Artemisia. See Table 2 for identity of population numbers. Populations 1 and 2 are A. campestris var. wormskioldii; populations 3 – 8 are A. campestris var. scouleriana; populations 9 and 10 are A. borealis.

1 2 3 4 5 6 7 8 9 10 1 2 0.944 3 0.982 0.933 4 0.956 0.992 0.951 5 0.988 0.967 0.981 0.970 6 0.981 0.974 0.968 0.979 0.988 7 0.948 0.943 0.940 0.953 0.946 0.973 8 0.938 0.908 0.932 0.924 0.919 0.944 0.988 9 0.969 0.909 0.970 0.939 0.959 0.948 0.915 0.911 10 0.940 0.952 0.935 0.969 0.954 0.955 0.917 0.890 0.971

NFGEL Artemisia Report

Figure 1. Left: Artemisia campestris var. scouleriana. Center: Artemisia campestris var. wormskioldii. Right: Artemisia campestris var. borealis. The scouleriana and wormskioldii plants both grew on the shore of the Columbia River. The borealis specimen is from Montana.

Specimens from the Oregon State University herbarium: scouleriana, Kemp 80099, Klickitat Co WA, 27 August 1980 wormskioldii, Henderson 5026, Sherman Co OR, 5 April 1925 borealis, Cronquist 7991, Carbon Co MT, 27 July 1955

NFGEL Artemisia Report

Figure 2. Left: Artemisia campestris var. scouleriana. Center: Artemisia campestris var. wormskioldii. Right: Artemisia campestris var. borealis. All photos at the same scale.

NFGEL Artemisia Report

Figure 3. Map of western Washington showing the locations of sampled populations. Population identities are identified by numbers (1 – 10). See Table 2 for identity of population numbers.

NFGEL Artemisia Report

A. campestris var scouleriana; Hanford White Bluff (#5)

A. campestris var scouleriana; Hanford Vernita Bridge (#6)

A. campestris var wormskioldii; Miller Island (#1)

A. campestris var scouleriana; Miller Island (#3)

A. campestris var wormskioldii; Beverly Island (#2)

A. campestris var scouleriana; Beverly Island (#4)

A. borealis; Buckhorn Mtn (#9)

A. borealis; Blue Mtns (#10)

A. campestris var scouleriana; Whidbey Island Fort Casey (#7)

A. campestris var scouleriana; Whidbey Island Ebey's Landing (#8)

0.035 0.030 0.025 0.020 0.015 0.010 0.005 0.000

Figure 4. Population phenogram of sampled Artemisia populations based on Nei’s genetic distances (using UPGMA in MEGA v. 6.0). All genetic distances are very low but there is structure among these very similar populations. See Table 2 for identity of population number.

NFGEL Artemisia Report

NEI (4) vs GGD (2) y = 1224.2x + 127.54 R² = 0.0769 350

300 250 200 150 Y GGD (2) 100 Linear (Y) 50 0 0.000 0.020 0.040 0.060 0.080 0.100 0.120 0.140 NEI (4)

Figure 5. Relationship between Nei’s genetic distance (x) and geographic distance (y) among sampled populations of borealis, scouleriana, and wormskioldii.

NEI vs GGD y = 2837x + 46.153 350.000 R² = 0.3748

300.000 250.000 200.000

GGD 150.000 Y

100.000 Linear (Y) 50.000 0.000 0.000 0.020 0.040 0.060 0.080 0.100 0.120 NEI

Figure 6. Relationship between Nei’s genetic distance (x) and geographic distance (y) among sampled populations of scouleriana and wormskioldii.

NFGEL Artemisia Report

Principal Coordinates

4 7 6 10 5

Coord.2 1 8 3

Coord. 1

Figure 7. Principal coordinates analysis (PCA) of populations of Artemisia sampled in this study. The first two axes, shown here, explain 53% and 33%, respectively, of the variation in the data. The third axis (not shown) explains an additional14%. Population numbers are identified in Table 2. Red symbols are A. campestris var. wormskioldii populations; blue are A. campestris var. scouleriana occurrences; yellow symbols indicate the A. borealis populations. Within the A. campestris symbols (blue and red), the square data points are from Miller Island; the two dashes are the Beverly sites; the diamonds are the Hanford populations; and the ‘x’ symbols are from Whidbey Island.

NFGEL Artemisia Report

A K=3

B K=4

Figure 8. Structure analysis of Artemisia reveal admixture and strong genetic grouping based on geographic population location. (A) Assignment of individual plants where each bar is one plant. The different colors indicate that proportion of a plants genotype assigned to a particular genetic cluster as depicted by alternate colors. Three genetic clusters are identified (K=3). (B) Four genetic clusters (K=4) are identified with assignments of individuals plants as vertical bars (as described in part (A)). (C) Tests to identify the most likely number of clusters. Left: greatest log-likelihood indicate K=3; Right: delta-K statistic indicates the most likely number of genetic clusters as 4 (K=4).

NFGEL Artemisia Report

APPENDIX A

Data Tables and Figure relevant to Artemisia

Table A1. Chromosome numbers reported for Artemisia borealis and A. campestris on IPCN. Table A2. Allele frequencies in populations of Artemisia campestris and A. borealis. Table A3. Percent polymorphic enzymes for borealis, scouleriana and wormskioldii. Table A4. Geographic distances in km among sampled populations of borealis, scouleriana, and wormskioldii. Table A5. Nei’s genetic distance among sampled populations of borealis, scouleriana, and wormskioldii. Table A6. Mantel test of relationship between geographic and genetic distances among sampled populations of borealis, scouleriana, and wormskioldii.

APPENDIX B

Keys relevant to Artemisia

Keys from two standard resources, modified (B. Wilson).

NFGEL Artemisia Report

Table A1. Chromosome numbers reported for Artemisia borealis and A. campestris on IPCN. Regions of origin are noted if obvious from the title.

Taxon 2n = Citation Region

A. borealis 18 Dalgaard 1988 Greenland A. borealis 36 Dawe & Murray 1979 A. borealis 36 Krogulevich 1976a Eurasia A. borealis 36 Krogulevich 1976b A. borealis 36 Krogulevich 1978 A. borealis 36 Krogulevich 1984 A. borealis 36 Lavrenko et al. 1989 Eurasia A. borealis 36 Murray & Kelso 1997 North America A. borealis 18 Volkova et al. 2003 Eurasia A. borealis 18 Zhukova & Petrovsky 1976 Eurasia A. borealis 36 Zhukova & Petrovsky 1980 Eurasia A. borealis 18 Zhukova 1980 Eurasia A. campestris 36 Alber & Pröbsting 1998 Eurasia A. campestris 36 Brañas & Xirau 1994. A. campestris 36 Javůrková-Jarolímová, V. 1992 Eurasia A. campestris 36 Kuzmanov et al. 1986 Eurasia A. campestris 36 Lavrenko & Serditov 1991 Eurasia A. campestris 18, 36 Montserrat et al. 2001 A. campestris 36 Morton 1981 North America A. campestris 18 Murin & Ferakova 1981 Eurasia A. campestris 18 Packer & Witkus 1982 A. campestris 36 Pogan et al. 1986 Eurasia A. campestris 18 Vallés & Torrel 1995 Mediterranean A. campestris 18 Xirau & Siljak-yakovlev 1997 Eurasia A. c. s. sericea 36 Kreitschitz & Vallés 2003 Eurasia A. c. v. campestris 36 Lövkist & Hultgård 1999 Eurasia A. c. v. glutinosa 18 Djerjour & Guittonneua 1976 A. c. v. glutinosa 18, 36 Ouyaya & Viano 1988 A. c. v. pacifica 18 Ward & Spellenberg 1986 North America

NFGEL Artemisia Report

References for chromosome counts from IPCN

Albers, F. & W. Pröbsting. 1998. In R. Wisskirchen & H. Haeupler, Standardliste der Farn- und Blütenpflanzen Deutschlands. Bundesamt für Naturschutz & Verlag Eugen Ulmer, Stuttgart. Brañas, M. O. & J. V. Xirau. 1994. Karyological studies in some taxa of the genus Artemisia (Asteraceae). Canad. J. Bot. 72: 1126–1135. Dalgaard, V. 1988. Chromosome numbers in some vascular plants from the Disko Bugt area (west Greenland). Willdenowia 18: 243–252. Dawe, J. C. & D. F. Murray. 1979. In IOPB chromosome number reports LXIII. Taxon 28: 265–268. Djerkjour, B., and G. G. Guittonneau. 1976. In IOPB chromosome number reports. LII. Taxon 25: 341–346. Javůrková-Jarolímová, V. 1992. In J. Měsíček & V. Javůrková-Jarolímová, List of Chromosome Numbers of the Czech Vascular Plants. Academia, Praha. Kreitschitz, A. & J. Vallès. 2003. New or rare data on chromosome numbers in several taxa of the genus Artemisia (Asteraceae) in Poland. Folia Geobot. 38(3): 333–343. Krogulevich, R. E. 1976a. Chromosome numbers of plant species from the Tunkinsky Alpes (East Sayan). News Sib. Depart. Acad. Sci. USSR, Ser. Biol. 15(3): 46–52. Krogulevich, R. E. 1976b. Rol poliploidii v genesise flory Putorana. In: L. I. Malyshev, ed. 1976 Flora Putorana: Materali k Poznaniiu Osobennostei Sostava i Genezisia Gornykh Subarkticheskikh Flor Sibir. Novosibirsk. Nauka, Novosibirsk Krogulevich, R. E. 1978. Kariologicheskij analiz vidov flory Vostochnogo Sajana. V Flora Pribajkal'ja. 19–48. Nauka, Novosibirsk. Krogulevich, R. E. 1978. Karyological analysis of the species of the flora of eastern Sayana. Pp. 19-48 in L. I. Malyshev & G. A. Peshlcova (eds.) Flora of the Prebaikal. Novosibirsk. Krogulevich, R. E. 1984. In R.E. Krogulevich & T.S. Rostovtseva, Khromosomnye Chisla Tsvetkovykh Rastenii i Sibiri Dal'nego Vostoka. Izdatel'stvo "Nauka", Sibirskoe Otdelenie, Novosibirsk. ý. Kuzmanov, B. A., S. B. Georgieva & V. A. Nikolova. 1986. Chromosome numbers of Bulgarian flowering plants. I. Fam. Asteraceae. Fitologija 31: 71–74. Lavrenko, A. N. & N. P. Serditov. 1991. Chromosome numbers in some plant species from the south-west of the Komi ASSR. Bot. Žhurn. (Moscow & Leningrad) 76: 769–771. Lavrenko, A. N., N. P. Serditov & Z. G. Ulle. 1989. Chromosome numbers in some species of flowering plants of the Urals (the Komi Autonomous Soviet Socialist Republic). Bot. Žhurn. (Moscow & Leningrad) 74: 1059–1061. Lövkvist, B. & U. M. Hultgård. 1999. Chromosome numbers in south Swedish vascular plants. Opera Bot. 137: 1–42. Montserrat, T., M. Cerbah, S. Siljak-yakovlev & J. Vallès. 2001. Étude cytogénétique de trois taxons du complexe d'Artemisia campestris L. (Asteraceae, Anthemideae): localisation de l'hétérochromatine et de l'ADN ribosomique. Bocconea 13: 623–628. Morton, J. K. 1981. Chromosome numbers in Compositae from Canada and the U.S.A. Bot. J. Linn. Soc. 82: 357–368. Murin, A. & V. Ferakova. 1981. Caryological study of Slovakian flora III. Acta Fac. Rerum Nat. Univ. Comenianae, Bot. 28: 59– 62. Murray, D. F. & S. Kelso. 1997. Chromosome numbers and notes on the taxonomy of selected Alaskan vascular plants. Rhodora 99: 33–55. Ouyahya, A. & J. Viano. 1988. Recherches cytogénétiques sur le genre Artemisia L. au Maroc. Bol. Soc. Brot., sér. 2 61: 105– 124. Packer, J. G. & R. Witkus. 1982. In IOPB chromosome number reports LXXV. Taxon 31: 363–364. Pogan, E., A. Jankun, J. Maslecka & H. Wcisslo. 1986. Further studies in chromosome numbers of Polish angiosperms. Part XIX. Acta Biol. Cracov., Ser. Bot. 28: 65–85. Vallés, J. & M. Torrell. 1995. Mediterranean chromosome number reports 5 (552--558). Fl. Medit. 5: 357–363. Volkova, S. A., E. V. Boyko & I. V. Antusheva. 2003. Chromosome numbers of some species of Kamchatka and Sakhalin flora. Bot. Žhurn. (Moscow & Leningrad) 88(9): 154–155. Ward, D. E. & R. W. Spellenberg. 1986. Chromosome counts of angiosperms of western North America. Phytologia 61: 119– 125. Xirau, J. V. & S. Siljak-yakovlev. 1997. Cytogenetic studies in the genus Artemisia L. (Asteraceae): fluorochrome-banded karyotypes of five taxa, including the Iberian endemic species Artemisia barrelieri Besser. Canad. J. Bot. 75: 595–606. Zhukova, P. G. & V. V. Petrovsky. 1976. Chromosome numbers of some Western Chukolka plant species, II. Bot. Žhurn. (Moscow & Leningrad) 61(7): 963–969. Zhukova, P. G. & V. V. Petrovsky. 1980. Chromosome numbers and taxonomy of some species of the Anyui Mts. Bot. Zhurn. 65 (5): 651–659. Zhukova, P. G. 1980. Chromosome numbers of some Southern Chukotka plant species. Bot. Zhurn. 65 (1): 51–59.

NFGEL Artemisia Report

Table A2. Allele frequencies in populations of Artemisia campestris and A. borealis. Populations are identified by taxon abbreviation and population number (see Table 2).

Allele Frequencies by Populations Taxon: A. c. w. A. c. w. A. c. s. A. c. s. A. c. s. A. c. s. A. c. s. A. c. s. A. b. A. b. Locus Allele 1 2 3 4 5 6 7 8 9 10 PGI-2 1 1.000 0.484 1.000 0.565 0.839 0.710 0.532 0.548 0.903 0.597 2 ----- 0.516 ----- 0.435 0.145 0.048 0.032 ----- 0.081 0.355 3 ------0.016 0.242 0.435 0.452 0.016 0.048 TPI 1 0.839 0.452 0.774 0.548 0.629 0.565 0.726 0.968 0.935 0.613 2 0.161 0.548 0.226 0.452 0.371 0.435 0.274 0.032 0.065 0.387 DIA 1 0.952 0.839 0.871 0.790 0.952 0.871 0.516 0.500 0.952 0.903 2 0.048 0.161 0.129 0.210 0.048 0.129 0.484 0.500 0.048 0.097 GLYDH 1 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 0.887 0.887 2 ------0.113 0.113 ME7 1 0.968 0.871 0.823 0.935 1.000 0.968 0.871 0.710 1.000 1.000 2 0.032 0.129 0.177 0.065 ----- 0.032 0.129 0.290 ------PGM 1 1.000 1.000 0.984 1.000 0.919 1.000 0.984 0.968 0.823 0.710 2 ------0.016 ----- 0.081 ----- 0.016 0.032 0.177 0.290 6PGD 1 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 0.742 0.710 2 ------0.258 0.290 SKD 1 1.000 1.000 0.694 1.000 0.823 0.984 0.968 0.984 0.839 0.952 2 ------0.306 ----- 0.177 0.016 0.032 0.016 0.161 0.048 ACO 1 0.694 0.694 0.839 0.855 0.661 0.742 0.710 0.726 1.000 1.000 2 0.306 0.258 0.129 0.048 0.339 0.258 0.290 0.274 ------3 ----- 0.048 0.032 0.097 ------

NFGEL Artemisia Report

Table A3. Percent polymorphic enzymes for Artemisia borealis (A. b.), Artemsia campestris var. scouleriana (A. c. s.), and Artemsia campestris var. wormskioldii (A. c. w.) samples. M = monomorphic. P = polymorphic.

Enzyme A. b. A. c. s. A. c. w.

6PGD P M M AAT-1 M M M AAT-2 M M M ACO M P P CAT M M M DIA P P P FEST M M M GLYDH P M M ME7 M P P PGI P P P SKD P P M TPI P P P UGPP P -- --

% 46(50)% 46% 38%

Table A4. Geographic distances in km among sampled populations of Artemisia borealis and A. campestris. See Table 2 for explanations of population number.

1 2 3 4 5 6 7 8 9 10 1 2 149.884 3 0.000 149.884 4 149.884 0.000 149.884 5 157.465 41.665 157.465 41.665 6 140.103 27.031 140.103 27.031 22.217 7 309.910 252.613 309.910 252.613 293.718 278.352 8 313.390 255.653 313.390 255.653 296.714 281.442 3.480 9 295.262 263.830 295.262 263.830 305.491 287.631 50.535 51.940 10 312.967 278.616 312.967 278.616 320.272 302.834 48.602 48.818 17.750

NFGEL Artemisia Report

Table A5. Nei’s genetic distance among sampled populations of Artemisia borealis and A. campestris. See Table 2 for explanations of population number.

1 2 3 4 5 6 7 8 9 10 1 2 0.057 3 0.019 0.069 4 0.045 0.008 0.050 5 0.012 0.033 0.019 0.031 6 0.019 0.026 0.032 0.021 0.012 7 0.054 0.058 0.062 0.048 0.056 0.027 8 0.065 0.096 0.071 0.079 0.085 0.057 0.012 9 0.031 0.096 0.030 0.063 0.042 0.054 0.089 0.093 10 0.062 0.049 0.067 0.032 0.047 0.046 0.087 0.117 0.030

Table A6. Mantel test of relationship between geographic and genetic distances among sampled populations of Artemisia borealis and A. campestris. 99 permutations in each test.

scouleriana + with A. borealis wormskioldii only No. samples 8 10 No. in matrix 28 45 SSx 0.016 0.031 SSy 349021.048 599426.172 SPxy 46.114 37.632 Rxy 0.612 0.277 Probability 0.040 0.020

NFGEL Artemisia Report

Appendix B. Keys from two standard resources, modified by B. Wilson to include keys to subtaxa or other relevant taxa mentioned, and to remove taxa irrelevant to the classification of scouleriana and wormskioldii. Material in bracket [] added to the key.

From: Shultz, Leila M. 2006. Artemisia. pp. 503-534 in Flora of North America Editorial Committee, eds., Flora of North America, volume 19, Magnoliophyta: Asteridae, part 6: Asteraceae, part 1. Oxford University Press, New York.

1. Plants 5-30(-80+) cm, often cespitose and/or mounded. 2. Perennials; leaves 2-3-palmately or –pinnnately lobed. 3. Leaves 2-3-pinnately or ternately lobed; corollas (at least lobes) usually yellow-orange or deep red; n latitudes and western mountains [involucres 3-4 x 3.5-4 mm] .... A. borealis 1’ Plants (10-)50-100 cm (not cespitose). 5. . . . leaves lobed, faces hairy. 6. Biennials or perennials, (10-)30-80(-150) cm; involucres 2-4.5(-7) mm diameter. 7. stems usually 1-5; heads in mostly leafless paniculiform arrays [involucres 2.5-3(-5) x 2-3.5(-7) mm] ...... A. campestris x. Perennials; stems 2-5; basal rosettes persistent ...... A. c. ssp. pacifica [including scouleriana] x. Biennials; stems 1(-3); basal rosettes not persistent (withering before flowering) y. Involucres globos, 35 x 3.5-5 mm; north of 50 degrees, primarily in Canada ...... A. c. ssp. canadenssi y. Involucres turbinate, 2-3 x 2-3 mm; south of 50 degrees, east from Rocky Mountains to coastal North America ...... A. c. ssp. caudata

From: Hitchcock, C. Leo, Arthur Cronquist, Marion Ownbey, and J. W. Thompson. 1955. Vascular Plants of the Pacific Northwest, Part 5: Compositae. Univerist of Washington Press, Seattle, Washington. x. Stems usually 1; basal rosettes commonly produced only during the first (non-flowering year and not persistent into the second year...... A. c. ssp. caudata x. Stems several; basal rosettes produced each year, these ordinarily peristent at least until anthesis ...... A. c. ssp. borealis 1. Low plants 1-4 dm tall, with narrow, spiciform-thyrsoid inflorescence and relatively large heads, the involucre commonly 3-4 mm high. 2. Herbage, especially the leaves, diensely and conspicuously sericeous; banks fo the Columbia ...... var. wormksioldii 2. Herbage moderately to scarcely sericeous; high mts. from > Wash. to N. Mont., and northward ...... var. borealis 1’ Taller plants, commonly 3-10 dm tall, with mostly larger and more open inflorescence, and generally with smaller heads, the involucre 2-3 mm high; widely scattered in the western cordillera at less alpine stations than var. borealis ...... var. scouleriana